blob: 3d9393b845b8f71095dd2b55a1bd601ef70c37cd [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001
Pekka J Enberg5ea626a2005-09-09 13:10:19 -07002 Overview of the Linux Virtual File System
Linus Torvalds1da177e2005-04-16 15:20:36 -07003
Pekka J Enberg5ea626a2005-09-09 13:10:19 -07004 Original author: Richard Gooch <rgooch@atnf.csiro.au>
Linus Torvalds1da177e2005-04-16 15:20:36 -07005
Borislav Petkov0746aec2007-07-15 23:41:19 -07006 Last updated on June 24, 2007.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -07007
8 Copyright (C) 1999 Richard Gooch
9 Copyright (C) 2005 Pekka Enberg
10
11 This file is released under the GPLv2.
Linus Torvalds1da177e2005-04-16 15:20:36 -070012
13
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080014Introduction
15============
Linus Torvalds1da177e2005-04-16 15:20:36 -070016
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080017The Virtual File System (also known as the Virtual Filesystem Switch)
18is the software layer in the kernel that provides the filesystem
19interface to userspace programs. It also provides an abstraction
20within the kernel which allows different filesystem implementations to
21coexist.
22
23VFS system calls open(2), stat(2), read(2), write(2), chmod(2) and so
24on are called from a process context. Filesystem locking is described
25in the document Documentation/filesystems/Locking.
Linus Torvalds1da177e2005-04-16 15:20:36 -070026
27
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080028Directory Entry Cache (dcache)
29------------------------------
Linus Torvalds1da177e2005-04-16 15:20:36 -070030
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080031The VFS implements the open(2), stat(2), chmod(2), and similar system
32calls. The pathname argument that is passed to them is used by the VFS
33to search through the directory entry cache (also known as the dentry
34cache or dcache). This provides a very fast look-up mechanism to
35translate a pathname (filename) into a specific dentry. Dentries live
36in RAM and are never saved to disc: they exist only for performance.
37
38The dentry cache is meant to be a view into your entire filespace. As
39most computers cannot fit all dentries in the RAM at the same time,
40some bits of the cache are missing. In order to resolve your pathname
41into a dentry, the VFS may have to resort to creating dentries along
42the way, and then loading the inode. This is done by looking up the
43inode.
Linus Torvalds1da177e2005-04-16 15:20:36 -070044
Pekka J Enberg5ea626a2005-09-09 13:10:19 -070045
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080046The Inode Object
47----------------
Linus Torvalds1da177e2005-04-16 15:20:36 -070048
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080049An individual dentry usually has a pointer to an inode. Inodes are
50filesystem objects such as regular files, directories, FIFOs and other
51beasts. They live either on the disc (for block device filesystems)
52or in the memory (for pseudo filesystems). Inodes that live on the
53disc are copied into the memory when required and changes to the inode
54are written back to disc. A single inode can be pointed to by multiple
55dentries (hard links, for example, do this).
Linus Torvalds1da177e2005-04-16 15:20:36 -070056
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080057To look up an inode requires that the VFS calls the lookup() method of
58the parent directory inode. This method is installed by the specific
59filesystem implementation that the inode lives in. Once the VFS has
60the required dentry (and hence the inode), we can do all those boring
61things like open(2) the file, or stat(2) it to peek at the inode
62data. The stat(2) operation is fairly simple: once the VFS has the
63dentry, it peeks at the inode data and passes some of it back to
64userspace.
Linus Torvalds1da177e2005-04-16 15:20:36 -070065
Linus Torvalds1da177e2005-04-16 15:20:36 -070066
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080067The File Object
68---------------
Linus Torvalds1da177e2005-04-16 15:20:36 -070069
70Opening a file requires another operation: allocation of a file
71structure (this is the kernel-side implementation of file
Pekka J Enberg5ea626a2005-09-09 13:10:19 -070072descriptors). The freshly allocated file structure is initialized with
Linus Torvalds1da177e2005-04-16 15:20:36 -070073a pointer to the dentry and a set of file operation member functions.
74These are taken from the inode data. The open() file method is then
Francis Galieguea33f3222010-04-23 00:08:02 +020075called so the specific filesystem implementation can do its work. You
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080076can see that this is another switch performed by the VFS. The file
77structure is placed into the file descriptor table for the process.
Linus Torvalds1da177e2005-04-16 15:20:36 -070078
79Reading, writing and closing files (and other assorted VFS operations)
80is done by using the userspace file descriptor to grab the appropriate
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080081file structure, and then calling the required file structure method to
82do whatever is required. For as long as the file is open, it keeps the
83dentry in use, which in turn means that the VFS inode is still in use.
Linus Torvalds1da177e2005-04-16 15:20:36 -070084
Pekka J Enberg5ea626a2005-09-09 13:10:19 -070085
86Registering and Mounting a Filesystem
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080087=====================================
Linus Torvalds1da177e2005-04-16 15:20:36 -070088
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080089To register and unregister a filesystem, use the following API
90functions:
Linus Torvalds1da177e2005-04-16 15:20:36 -070091
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080092 #include <linux/fs.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -070093
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080094 extern int register_filesystem(struct file_system_type *);
95 extern int unregister_filesystem(struct file_system_type *);
Linus Torvalds1da177e2005-04-16 15:20:36 -070096
Pekka Enbergcc7d1f82005-11-07 01:01:08 -080097The passed struct file_system_type describes your filesystem. When a
Al Viro1a102ff2011-03-16 09:07:58 -040098request is made to mount a filesystem onto a directory in your namespace,
99the VFS will call the appropriate mount() method for the specific
Lucas De Marchi25985ed2011-03-30 22:57:33 -0300100filesystem. New vfsmount referring to the tree returned by ->mount()
Al Viro1a102ff2011-03-16 09:07:58 -0400101will be attached to the mountpoint, so that when pathname resolution
102reaches the mountpoint it will jump into the root of that vfsmount.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700103
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800104You can see all filesystems that are registered to the kernel in the
105file /proc/filesystems.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700106
107
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700108struct file_system_type
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800109-----------------------
Linus Torvalds1da177e2005-04-16 15:20:36 -0700110
Al Viro1a102ff2011-03-16 09:07:58 -0400111This describes the filesystem. As of kernel 2.6.39, the following
Linus Torvalds1da177e2005-04-16 15:20:36 -0700112members are defined:
113
114struct file_system_type {
115 const char *name;
116 int fs_flags;
Al Viro1a102ff2011-03-16 09:07:58 -0400117 struct dentry (*mount) (struct file_system_type *, int,
118 const char *, void *);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700119 void (*kill_sb) (struct super_block *);
120 struct module *owner;
121 struct file_system_type * next;
122 struct list_head fs_supers;
Borislav Petkov0746aec2007-07-15 23:41:19 -0700123 struct lock_class_key s_lock_key;
124 struct lock_class_key s_umount_key;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700125};
126
127 name: the name of the filesystem type, such as "ext2", "iso9660",
128 "msdos" and so on
129
130 fs_flags: various flags (i.e. FS_REQUIRES_DEV, FS_NO_DCACHE, etc.)
131
Al Viro1a102ff2011-03-16 09:07:58 -0400132 mount: the method to call when a new instance of this
Linus Torvalds1da177e2005-04-16 15:20:36 -0700133 filesystem should be mounted
134
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700135 kill_sb: the method to call when an instance of this filesystem
Al Viro1a102ff2011-03-16 09:07:58 -0400136 should be shut down
Linus Torvalds1da177e2005-04-16 15:20:36 -0700137
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700138 owner: for internal VFS use: you should initialize this to THIS_MODULE in
139 most cases.
140
141 next: for internal VFS use: you should initialize this to NULL
142
Borislav Petkov0746aec2007-07-15 23:41:19 -0700143 s_lock_key, s_umount_key: lockdep-specific
144
Al Viro1a102ff2011-03-16 09:07:58 -0400145The mount() method has the following arguments:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700146
Matt LaPlanted9195882008-07-25 19:45:33 -0700147 struct file_system_type *fs_type: describes the filesystem, partly initialized
Borislav Petkov0746aec2007-07-15 23:41:19 -0700148 by the specific filesystem code
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700149
150 int flags: mount flags
151
152 const char *dev_name: the device name we are mounting.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700153
154 void *data: arbitrary mount options, usually comes as an ASCII
Miklos Szeredif84e3f52008-02-08 04:21:34 -0800155 string (see "Mount Options" section)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700156
Al Viro1a102ff2011-03-16 09:07:58 -0400157The mount() method must return the root dentry of the tree requested by
158caller. An active reference to its superblock must be grabbed and the
159superblock must be locked. On failure it should return ERR_PTR(error).
Linus Torvalds1da177e2005-04-16 15:20:36 -0700160
Al Viro1a102ff2011-03-16 09:07:58 -0400161The arguments match those of mount(2) and their interpretation
162depends on filesystem type. E.g. for block filesystems, dev_name is
163interpreted as block device name, that device is opened and if it
164contains a suitable filesystem image the method creates and initializes
165struct super_block accordingly, returning its root dentry to caller.
166
167->mount() may choose to return a subtree of existing filesystem - it
168doesn't have to create a new one. The main result from the caller's
169point of view is a reference to dentry at the root of (sub)tree to
170be attached; creation of new superblock is a common side effect.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700171
172The most interesting member of the superblock structure that the
Al Viro1a102ff2011-03-16 09:07:58 -0400173mount() method fills in is the "s_op" field. This is a pointer to
Linus Torvalds1da177e2005-04-16 15:20:36 -0700174a "struct super_operations" which describes the next level of the
175filesystem implementation.
176
Al Viro1a102ff2011-03-16 09:07:58 -0400177Usually, a filesystem uses one of the generic mount() implementations
178and provides a fill_super() callback instead. The generic variants are:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700179
Al Viro1a102ff2011-03-16 09:07:58 -0400180 mount_bdev: mount a filesystem residing on a block device
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700181
Al Viro1a102ff2011-03-16 09:07:58 -0400182 mount_nodev: mount a filesystem that is not backed by a device
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700183
Al Viro1a102ff2011-03-16 09:07:58 -0400184 mount_single: mount a filesystem which shares the instance between
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700185 all mounts
186
Al Viro1a102ff2011-03-16 09:07:58 -0400187A fill_super() callback implementation has the following arguments:
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700188
Al Viro1a102ff2011-03-16 09:07:58 -0400189 struct super_block *sb: the superblock structure. The callback
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700190 must initialize this properly.
191
192 void *data: arbitrary mount options, usually comes as an ASCII
Miklos Szeredif84e3f52008-02-08 04:21:34 -0800193 string (see "Mount Options" section)
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700194
195 int silent: whether or not to be silent on error
196
197
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800198The Superblock Object
199=====================
200
201A superblock object represents a mounted filesystem.
202
203
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700204struct super_operations
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800205-----------------------
Linus Torvalds1da177e2005-04-16 15:20:36 -0700206
207This describes how the VFS can manipulate the superblock of your
Borislav Petkov422b14c2007-07-15 23:41:43 -0700208filesystem. As of kernel 2.6.22, the following members are defined:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700209
210struct super_operations {
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700211 struct inode *(*alloc_inode)(struct super_block *sb);
212 void (*destroy_inode)(struct inode *);
213
Christoph Hellwigaa385722011-05-27 06:53:02 -0400214 void (*dirty_inode) (struct inode *, int flags);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700215 int (*write_inode) (struct inode *, int);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700216 void (*drop_inode) (struct inode *);
217 void (*delete_inode) (struct inode *);
218 void (*put_super) (struct super_block *);
219 void (*write_super) (struct super_block *);
220 int (*sync_fs)(struct super_block *sb, int wait);
Takashi Satoc4be0c12009-01-09 16:40:58 -0800221 int (*freeze_fs) (struct super_block *);
222 int (*unfreeze_fs) (struct super_block *);
David Howells726c3342006-06-23 02:02:58 -0700223 int (*statfs) (struct dentry *, struct kstatfs *);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700224 int (*remount_fs) (struct super_block *, int *, char *);
225 void (*clear_inode) (struct inode *);
226 void (*umount_begin) (struct super_block *);
227
Al Viro34c80b12011-12-08 21:32:45 -0500228 int (*show_options)(struct seq_file *, struct dentry *);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700229
230 ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
231 ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
Dave Chinner0e1fdaf2011-07-08 14:14:44 +1000232 int (*nr_cached_objects)(struct super_block *);
233 void (*free_cached_objects)(struct super_block *, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700234};
235
236All methods are called without any locks being held, unless otherwise
237noted. This means that most methods can block safely. All methods are
238only called from a process context (i.e. not from an interrupt handler
239or bottom half).
240
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700241 alloc_inode: this method is called by inode_alloc() to allocate memory
NeilBrown341546f2006-03-25 03:07:56 -0800242 for struct inode and initialize it. If this function is not
243 defined, a simple 'struct inode' is allocated. Normally
244 alloc_inode will be used to allocate a larger structure which
245 contains a 'struct inode' embedded within it.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700246
247 destroy_inode: this method is called by destroy_inode() to release
NeilBrown341546f2006-03-25 03:07:56 -0800248 resources allocated for struct inode. It is only required if
249 ->alloc_inode was defined and simply undoes anything done by
250 ->alloc_inode.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700251
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700252 dirty_inode: this method is called by the VFS to mark an inode dirty.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700253
254 write_inode: this method is called when the VFS needs to write an
255 inode to disc. The second parameter indicates whether the write
256 should be synchronous or not, not all filesystems check this flag.
257
Linus Torvalds1da177e2005-04-16 15:20:36 -0700258 drop_inode: called when the last access to the inode is dropped,
Dave Chinnerf283c862011-03-22 22:23:39 +1100259 with the inode->i_lock spinlock held.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700260
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700261 This method should be either NULL (normal UNIX filesystem
Linus Torvalds1da177e2005-04-16 15:20:36 -0700262 semantics) or "generic_delete_inode" (for filesystems that do not
263 want to cache inodes - causing "delete_inode" to always be
264 called regardless of the value of i_nlink)
265
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700266 The "generic_delete_inode()" behavior is equivalent to the
Linus Torvalds1da177e2005-04-16 15:20:36 -0700267 old practice of using "force_delete" in the put_inode() case,
268 but does not have the races that the "force_delete()" approach
269 had.
270
271 delete_inode: called when the VFS wants to delete an inode
272
Linus Torvalds1da177e2005-04-16 15:20:36 -0700273 put_super: called when the VFS wishes to free the superblock
274 (i.e. unmount). This is called with the superblock lock held
275
276 write_super: called when the VFS superblock needs to be written to
277 disc. This method is optional
278
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700279 sync_fs: called when VFS is writing out all dirty data associated with
280 a superblock. The second parameter indicates whether the method
281 should wait until the write out has been completed. Optional.
282
Takashi Satoc4be0c12009-01-09 16:40:58 -0800283 freeze_fs: called when VFS is locking a filesystem and
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800284 forcing it into a consistent state. This method is currently
285 used by the Logical Volume Manager (LVM).
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700286
Takashi Satoc4be0c12009-01-09 16:40:58 -0800287 unfreeze_fs: called when VFS is unlocking a filesystem and making it writable
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700288 again.
289
Adrian McMenamin66672fe2009-04-20 18:38:28 -0700290 statfs: called when the VFS needs to get filesystem statistics.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700291
292 remount_fs: called when the filesystem is remounted. This is called
293 with the kernel lock held
294
295 clear_inode: called then the VFS clears the inode. Optional
296
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700297 umount_begin: called when the VFS is unmounting a filesystem.
298
Miklos Szeredif84e3f52008-02-08 04:21:34 -0800299 show_options: called by the VFS to show mount options for
300 /proc/<pid>/mounts. (see "Mount Options" section)
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700301
302 quota_read: called by the VFS to read from filesystem quota file.
303
304 quota_write: called by the VFS to write to filesystem quota file.
305
Dave Chinner0e1fdaf2011-07-08 14:14:44 +1000306 nr_cached_objects: called by the sb cache shrinking function for the
307 filesystem to return the number of freeable cached objects it contains.
308 Optional.
309
310 free_cache_objects: called by the sb cache shrinking function for the
311 filesystem to scan the number of objects indicated to try to free them.
312 Optional, but any filesystem implementing this method needs to also
313 implement ->nr_cached_objects for it to be called correctly.
314
315 We can't do anything with any errors that the filesystem might
316 encountered, hence the void return type. This will never be called if
317 the VM is trying to reclaim under GFP_NOFS conditions, hence this
318 method does not need to handle that situation itself.
319
Dave Chinner8ab47662011-07-08 14:14:45 +1000320 Implementations must include conditional reschedule calls inside any
321 scanning loop that is done. This allows the VFS to determine
322 appropriate scan batch sizes without having to worry about whether
323 implementations will cause holdoff problems due to large scan batch
324 sizes.
325
David Howells12debc42008-02-07 00:15:52 -0800326Whoever sets up the inode is responsible for filling in the "i_op" field. This
327is a pointer to a "struct inode_operations" which describes the methods that
328can be performed on individual inodes.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700329
330
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800331The Inode Object
332================
333
334An inode object represents an object within the filesystem.
335
336
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700337struct inode_operations
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800338-----------------------
Linus Torvalds1da177e2005-04-16 15:20:36 -0700339
340This describes how the VFS can manipulate an inode in your
Borislav Petkov422b14c2007-07-15 23:41:43 -0700341filesystem. As of kernel 2.6.22, the following members are defined:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700342
343struct inode_operations {
Al Viro4acdaf22011-07-26 01:42:34 -0400344 int (*create) (struct inode *,struct dentry *, umode_t, struct nameidata *);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700345 struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700346 int (*link) (struct dentry *,struct inode *,struct dentry *);
347 int (*unlink) (struct inode *,struct dentry *);
348 int (*symlink) (struct inode *,struct dentry *,const char *);
Al Viro18bb1db2011-07-26 01:41:39 -0400349 int (*mkdir) (struct inode *,struct dentry *,umode_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700350 int (*rmdir) (struct inode *,struct dentry *);
Al Viro1a67aaf2011-07-26 01:52:52 -0400351 int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700352 int (*rename) (struct inode *, struct dentry *,
353 struct inode *, struct dentry *);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700354 int (*readlink) (struct dentry *, char __user *,int);
355 void * (*follow_link) (struct dentry *, struct nameidata *);
356 void (*put_link) (struct dentry *, struct nameidata *, void *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700357 void (*truncate) (struct inode *);
Al Viro10556cb2011-06-20 19:28:19 -0400358 int (*permission) (struct inode *, int);
Christoph Hellwig4e34e712011-07-23 17:37:31 +0200359 int (*get_acl)(struct inode *, int);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700360 int (*setattr) (struct dentry *, struct iattr *);
361 int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
362 int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
363 ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
364 ssize_t (*listxattr) (struct dentry *, char *, size_t);
365 int (*removexattr) (struct dentry *, const char *);
Borislav Petkov422b14c2007-07-15 23:41:43 -0700366 void (*truncate_range)(struct inode *, loff_t, loff_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700367};
368
369Again, all methods are called without any locks being held, unless
370otherwise noted.
371
Linus Torvalds1da177e2005-04-16 15:20:36 -0700372 create: called by the open(2) and creat(2) system calls. Only
373 required if you want to support regular files. The dentry you
374 get should not have an inode (i.e. it should be a negative
375 dentry). Here you will probably call d_instantiate() with the
376 dentry and the newly created inode
377
378 lookup: called when the VFS needs to look up an inode in a parent
379 directory. The name to look for is found in the dentry. This
380 method must call d_add() to insert the found inode into the
381 dentry. The "i_count" field in the inode structure should be
382 incremented. If the named inode does not exist a NULL inode
383 should be inserted into the dentry (this is called a negative
384 dentry). Returning an error code from this routine must only
385 be done on a real error, otherwise creating inodes with system
386 calls like create(2), mknod(2), mkdir(2) and so on will fail.
387 If you wish to overload the dentry methods then you should
388 initialise the "d_dop" field in the dentry; this is a pointer
389 to a struct "dentry_operations".
390 This method is called with the directory inode semaphore held
391
392 link: called by the link(2) system call. Only required if you want
393 to support hard links. You will probably need to call
394 d_instantiate() just as you would in the create() method
395
396 unlink: called by the unlink(2) system call. Only required if you
397 want to support deleting inodes
398
399 symlink: called by the symlink(2) system call. Only required if you
400 want to support symlinks. You will probably need to call
401 d_instantiate() just as you would in the create() method
402
403 mkdir: called by the mkdir(2) system call. Only required if you want
404 to support creating subdirectories. You will probably need to
405 call d_instantiate() just as you would in the create() method
406
407 rmdir: called by the rmdir(2) system call. Only required if you want
408 to support deleting subdirectories
409
410 mknod: called by the mknod(2) system call to create a device (char,
411 block) inode or a named pipe (FIFO) or socket. Only required
412 if you want to support creating these types of inodes. You
413 will probably need to call d_instantiate() just as you would
414 in the create() method
415
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800416 rename: called by the rename(2) system call to rename the object to
417 have the parent and name given by the second inode and dentry.
418
Linus Torvalds1da177e2005-04-16 15:20:36 -0700419 readlink: called by the readlink(2) system call. Only required if
420 you want to support reading symbolic links
421
422 follow_link: called by the VFS to follow a symbolic link to the
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700423 inode it points to. Only required if you want to support
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800424 symbolic links. This method returns a void pointer cookie
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700425 that is passed to put_link().
426
427 put_link: called by the VFS to release resources allocated by
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800428 follow_link(). The cookie returned by follow_link() is passed
Paolo Ornati670e9f32006-10-03 22:57:56 +0200429 to this method as the last parameter. It is used by
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800430 filesystems such as NFS where page cache is not stable
431 (i.e. page that was installed when the symbolic link walk
432 started might not be in the page cache at the end of the
433 walk).
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700434
npiggin@suse.de7bb46a62010-05-27 01:05:33 +1000435 truncate: Deprecated. This will not be called if ->setsize is defined.
436 Called by the VFS to change the size of a file. The
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800437 i_size field of the inode is set to the desired size by the
438 VFS before this method is called. This method is called by
439 the truncate(2) system call and related functionality.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700440
npiggin@suse.de7bb46a62010-05-27 01:05:33 +1000441 Note: ->truncate and vmtruncate are deprecated. Do not add new
442 instances/calls of these. Filesystems should be converted to do their
443 truncate sequence via ->setattr().
444
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700445 permission: called by the VFS to check for access rights on a POSIX-like
446 filesystem.
447
Al Viro10556cb2011-06-20 19:28:19 -0400448 May be called in rcu-walk mode (mask & MAY_NOT_BLOCK). If in rcu-walk
Nick Piggina82416d2011-01-14 02:26:53 +0000449 mode, the filesystem must check the permission without blocking or
Nick Pigginb74c79e2011-01-07 17:49:58 +1100450 storing to the inode.
451
452 If a situation is encountered that rcu-walk cannot handle, return
453 -ECHILD and it will be called again in ref-walk mode.
454
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800455 setattr: called by the VFS to set attributes for a file. This method
456 is called by chmod(2) and related system calls.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700457
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800458 getattr: called by the VFS to get attributes of a file. This method
459 is called by stat(2) and related system calls.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700460
461 setxattr: called by the VFS to set an extended attribute for a file.
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800462 Extended attribute is a name:value pair associated with an
463 inode. This method is called by setxattr(2) system call.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700464
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800465 getxattr: called by the VFS to retrieve the value of an extended
466 attribute name. This method is called by getxattr(2) function
467 call.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700468
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800469 listxattr: called by the VFS to list all extended attributes for a
470 given file. This method is called by listxattr(2) system call.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700471
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800472 removexattr: called by the VFS to remove an extended attribute from
473 a file. This method is called by removexattr(2) system call.
474
Borislav Petkov422b14c2007-07-15 23:41:43 -0700475 truncate_range: a method provided by the underlying filesystem to truncate a
476 range of blocks , i.e. punch a hole somewhere in a file.
477
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800478
479The Address Space Object
480========================
481
NeilBrown341546f2006-03-25 03:07:56 -0800482The address space object is used to group and manage pages in the page
483cache. It can be used to keep track of the pages in a file (or
484anything else) and also track the mapping of sections of the file into
485process address spaces.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700486
NeilBrown341546f2006-03-25 03:07:56 -0800487There are a number of distinct yet related services that an
488address-space can provide. These include communicating memory
489pressure, page lookup by address, and keeping track of pages tagged as
490Dirty or Writeback.
491
NeilBrowna9e102b2006-03-25 03:08:29 -0800492The first can be used independently to the others. The VM can try to
NeilBrown341546f2006-03-25 03:07:56 -0800493either write dirty pages in order to clean them, or release clean
494pages in order to reuse them. To do this it can call the ->writepage
495method on dirty pages, and ->releasepage on clean pages with
496PagePrivate set. Clean pages without PagePrivate and with no external
497references will be released without notice being given to the
498address_space.
499
NeilBrowna9e102b2006-03-25 03:08:29 -0800500To achieve this functionality, pages need to be placed on an LRU with
NeilBrown341546f2006-03-25 03:07:56 -0800501lru_cache_add and mark_page_active needs to be called whenever the
502page is used.
503
504Pages are normally kept in a radix tree index by ->index. This tree
505maintains information about the PG_Dirty and PG_Writeback status of
506each page, so that pages with either of these flags can be found
507quickly.
508
509The Dirty tag is primarily used by mpage_writepages - the default
510->writepages method. It uses the tag to find dirty pages to call
511->writepage on. If mpage_writepages is not used (i.e. the address
NeilBrowna9e102b2006-03-25 03:08:29 -0800512provides its own ->writepages) , the PAGECACHE_TAG_DIRTY tag is
NeilBrown341546f2006-03-25 03:07:56 -0800513almost unused. write_inode_now and sync_inode do use it (through
514__sync_single_inode) to check if ->writepages has been successful in
515writing out the whole address_space.
516
517The Writeback tag is used by filemap*wait* and sync_page* functions,
Christoph Hellwig94004ed2009-09-30 22:16:33 +0200518via filemap_fdatawait_range, to wait for all writeback to
NeilBrown341546f2006-03-25 03:07:56 -0800519complete. While waiting ->sync_page (if defined) will be called on
NeilBrowna9e102b2006-03-25 03:08:29 -0800520each page that is found to require writeback.
NeilBrown341546f2006-03-25 03:07:56 -0800521
522An address_space handler may attach extra information to a page,
523typically using the 'private' field in the 'struct page'. If such
524information is attached, the PG_Private flag should be set. This will
NeilBrowna9e102b2006-03-25 03:08:29 -0800525cause various VM routines to make extra calls into the address_space
NeilBrown341546f2006-03-25 03:07:56 -0800526handler to deal with that data.
527
528An address space acts as an intermediate between storage and
529application. Data is read into the address space a whole page at a
530time, and provided to the application either by copying of the page,
531or by memory-mapping the page.
532Data is written into the address space by the application, and then
533written-back to storage typically in whole pages, however the
NeilBrowna9e102b2006-03-25 03:08:29 -0800534address_space has finer control of write sizes.
NeilBrown341546f2006-03-25 03:07:56 -0800535
536The read process essentially only requires 'readpage'. The write
Nick Piggin4e02ed42008-10-29 14:00:55 -0700537process is more complicated and uses write_begin/write_end or
NeilBrown341546f2006-03-25 03:07:56 -0800538set_page_dirty to write data into the address_space, and writepage,
539sync_page, and writepages to writeback data to storage.
540
541Adding and removing pages to/from an address_space is protected by the
542inode's i_mutex.
543
544When data is written to a page, the PG_Dirty flag should be set. It
545typically remains set until writepage asks for it to be written. This
546should clear PG_Dirty and set PG_Writeback. It can be actually
547written at any point after PG_Dirty is clear. Once it is known to be
548safe, PG_Writeback is cleared.
549
550Writeback makes use of a writeback_control structure...
Linus Torvalds1da177e2005-04-16 15:20:36 -0700551
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700552struct address_space_operations
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800553-------------------------------
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700554
555This describes how the VFS can manipulate mapping of a file to page cache in
Borislav Petkov422b14c2007-07-15 23:41:43 -0700556your filesystem. As of kernel 2.6.22, the following members are defined:
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700557
558struct address_space_operations {
559 int (*writepage)(struct page *page, struct writeback_control *wbc);
560 int (*readpage)(struct file *, struct page *);
561 int (*sync_page)(struct page *);
562 int (*writepages)(struct address_space *, struct writeback_control *);
563 int (*set_page_dirty)(struct page *page);
564 int (*readpages)(struct file *filp, struct address_space *mapping,
565 struct list_head *pages, unsigned nr_pages);
Nick Pigginafddba42007-10-16 01:25:01 -0700566 int (*write_begin)(struct file *, struct address_space *mapping,
567 loff_t pos, unsigned len, unsigned flags,
568 struct page **pagep, void **fsdata);
569 int (*write_end)(struct file *, struct address_space *mapping,
570 loff_t pos, unsigned len, unsigned copied,
571 struct page *page, void *fsdata);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700572 sector_t (*bmap)(struct address_space *, sector_t);
573 int (*invalidatepage) (struct page *, unsigned long);
574 int (*releasepage) (struct page *, int);
Linus Torvalds6072d132010-12-01 13:35:19 -0500575 void (*freepage)(struct page *);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700576 ssize_t (*direct_IO)(int, struct kiocb *, const struct iovec *iov,
577 loff_t offset, unsigned long nr_segs);
578 struct page* (*get_xip_page)(struct address_space *, sector_t,
579 int);
NeilBrown341546f2006-03-25 03:07:56 -0800580 /* migrate the contents of a page to the specified target */
581 int (*migratepage) (struct page *, struct page *);
Borislav Petkov422b14c2007-07-15 23:41:43 -0700582 int (*launder_page) (struct page *);
Andi Kleen25718732009-09-16 11:50:13 +0200583 int (*error_remove_page) (struct mapping *mapping, struct page *page);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700584};
585
NeilBrown341546f2006-03-25 03:07:56 -0800586 writepage: called by the VM to write a dirty page to backing store.
NeilBrowna9e102b2006-03-25 03:08:29 -0800587 This may happen for data integrity reasons (i.e. 'sync'), or
NeilBrown341546f2006-03-25 03:07:56 -0800588 to free up memory (flush). The difference can be seen in
589 wbc->sync_mode.
590 The PG_Dirty flag has been cleared and PageLocked is true.
591 writepage should start writeout, should set PG_Writeback,
592 and should make sure the page is unlocked, either synchronously
593 or asynchronously when the write operation completes.
594
595 If wbc->sync_mode is WB_SYNC_NONE, ->writepage doesn't have to
NeilBrowna9e102b2006-03-25 03:08:29 -0800596 try too hard if there are problems, and may choose to write out
597 other pages from the mapping if that is easier (e.g. due to
598 internal dependencies). If it chooses not to start writeout, it
599 should return AOP_WRITEPAGE_ACTIVATE so that the VM will not keep
NeilBrown341546f2006-03-25 03:07:56 -0800600 calling ->writepage on that page.
601
602 See the file "Locking" for more details.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700603
604 readpage: called by the VM to read a page from backing store.
NeilBrown341546f2006-03-25 03:07:56 -0800605 The page will be Locked when readpage is called, and should be
606 unlocked and marked uptodate once the read completes.
607 If ->readpage discovers that it needs to unlock the page for
608 some reason, it can do so, and then return AOP_TRUNCATED_PAGE.
NeilBrowna9e102b2006-03-25 03:08:29 -0800609 In this case, the page will be relocated, relocked and if
NeilBrown341546f2006-03-25 03:07:56 -0800610 that all succeeds, ->readpage will be called again.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700611
612 sync_page: called by the VM to notify the backing store to perform all
613 queued I/O operations for a page. I/O operations for other pages
614 associated with this address_space object may also be performed.
615
NeilBrown341546f2006-03-25 03:07:56 -0800616 This function is optional and is called only for pages with
617 PG_Writeback set while waiting for the writeback to complete.
618
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700619 writepages: called by the VM to write out pages associated with the
NeilBrowna9e102b2006-03-25 03:08:29 -0800620 address_space object. If wbc->sync_mode is WBC_SYNC_ALL, then
621 the writeback_control will specify a range of pages that must be
622 written out. If it is WBC_SYNC_NONE, then a nr_to_write is given
NeilBrown341546f2006-03-25 03:07:56 -0800623 and that many pages should be written if possible.
624 If no ->writepages is given, then mpage_writepages is used
NeilBrowna9e102b2006-03-25 03:08:29 -0800625 instead. This will choose pages from the address space that are
NeilBrown341546f2006-03-25 03:07:56 -0800626 tagged as DIRTY and will pass them to ->writepage.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700627
628 set_page_dirty: called by the VM to set a page dirty.
NeilBrown341546f2006-03-25 03:07:56 -0800629 This is particularly needed if an address space attaches
630 private data to a page, and that data needs to be updated when
631 a page is dirtied. This is called, for example, when a memory
632 mapped page gets modified.
633 If defined, it should set the PageDirty flag, and the
634 PAGECACHE_TAG_DIRTY tag in the radix tree.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700635
636 readpages: called by the VM to read pages associated with the address_space
NeilBrown341546f2006-03-25 03:07:56 -0800637 object. This is essentially just a vector version of
638 readpage. Instead of just one page, several pages are
639 requested.
NeilBrowna9e102b2006-03-25 03:08:29 -0800640 readpages is only used for read-ahead, so read errors are
NeilBrown341546f2006-03-25 03:07:56 -0800641 ignored. If anything goes wrong, feel free to give up.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700642
Nick Piggin4e02ed42008-10-29 14:00:55 -0700643 write_begin:
Nick Pigginafddba42007-10-16 01:25:01 -0700644 Called by the generic buffered write code to ask the filesystem to
645 prepare to write len bytes at the given offset in the file. The
646 address_space should check that the write will be able to complete,
647 by allocating space if necessary and doing any other internal
648 housekeeping. If the write will update parts of any basic-blocks on
649 storage, then those blocks should be pre-read (if they haven't been
650 read already) so that the updated blocks can be written out properly.
651
652 The filesystem must return the locked pagecache page for the specified
653 offset, in *pagep, for the caller to write into.
654
Nick Piggin4e02ed42008-10-29 14:00:55 -0700655 It must be able to cope with short writes (where the length passed to
656 write_begin is greater than the number of bytes copied into the page).
657
Nick Pigginafddba42007-10-16 01:25:01 -0700658 flags is a field for AOP_FLAG_xxx flags, described in
659 include/linux/fs.h.
660
661 A void * may be returned in fsdata, which then gets passed into
662 write_end.
663
664 Returns 0 on success; < 0 on failure (which is the error code), in
665 which case write_end is not called.
666
667 write_end: After a successful write_begin, and data copy, write_end must
668 be called. len is the original len passed to write_begin, and copied
669 is the amount that was able to be copied (copied == len is always true
670 if write_begin was called with the AOP_FLAG_UNINTERRUPTIBLE flag).
671
672 The filesystem must take care of unlocking the page and releasing it
673 refcount, and updating i_size.
674
675 Returns < 0 on failure, otherwise the number of bytes (<= 'copied')
676 that were able to be copied into pagecache.
677
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700678 bmap: called by the VFS to map a logical block offset within object to
NeilBrowna9e102b2006-03-25 03:08:29 -0800679 physical block number. This method is used by the FIBMAP
NeilBrown341546f2006-03-25 03:07:56 -0800680 ioctl and for working with swap-files. To be able to swap to
NeilBrowna9e102b2006-03-25 03:08:29 -0800681 a file, the file must have a stable mapping to a block
NeilBrown341546f2006-03-25 03:07:56 -0800682 device. The swap system does not go through the filesystem
683 but instead uses bmap to find out where the blocks in the file
684 are and uses those addresses directly.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700685
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700686
NeilBrown341546f2006-03-25 03:07:56 -0800687 invalidatepage: If a page has PagePrivate set, then invalidatepage
688 will be called when part or all of the page is to be removed
NeilBrowna9e102b2006-03-25 03:08:29 -0800689 from the address space. This generally corresponds to either a
NeilBrown341546f2006-03-25 03:07:56 -0800690 truncation or a complete invalidation of the address space
691 (in the latter case 'offset' will always be 0).
692 Any private data associated with the page should be updated
693 to reflect this truncation. If offset is 0, then
694 the private data should be released, because the page
695 must be able to be completely discarded. This may be done by
696 calling the ->releasepage function, but in this case the
697 release MUST succeed.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700698
NeilBrown341546f2006-03-25 03:07:56 -0800699 releasepage: releasepage is called on PagePrivate pages to indicate
700 that the page should be freed if possible. ->releasepage
701 should remove any private data from the page and clear the
Andrew Morton4fe65ca2010-12-02 14:31:19 -0800702 PagePrivate flag. If releasepage() fails for some reason, it must
703 indicate failure with a 0 return value.
704 releasepage() is used in two distinct though related cases. The
705 first is when the VM finds a clean page with no active users and
NeilBrown341546f2006-03-25 03:07:56 -0800706 wants to make it a free page. If ->releasepage succeeds, the
707 page will be removed from the address_space and become free.
708
Shaun Zinckbc5b1d52007-10-20 02:35:36 +0200709 The second case is when a request has been made to invalidate
NeilBrown341546f2006-03-25 03:07:56 -0800710 some or all pages in an address_space. This can happen
711 through the fadvice(POSIX_FADV_DONTNEED) system call or by the
712 filesystem explicitly requesting it as nfs and 9fs do (when
713 they believe the cache may be out of date with storage) by
714 calling invalidate_inode_pages2().
715 If the filesystem makes such a call, and needs to be certain
NeilBrowna9e102b2006-03-25 03:08:29 -0800716 that all pages are invalidated, then its releasepage will
NeilBrown341546f2006-03-25 03:07:56 -0800717 need to ensure this. Possibly it can clear the PageUptodate
718 bit if it cannot free private data yet.
719
Linus Torvalds6072d132010-12-01 13:35:19 -0500720 freepage: freepage is called once the page is no longer visible in
721 the page cache in order to allow the cleanup of any private
722 data. Since it may be called by the memory reclaimer, it
723 should not assume that the original address_space mapping still
724 exists, and it should not block.
725
NeilBrown341546f2006-03-25 03:07:56 -0800726 direct_IO: called by the generic read/write routines to perform
727 direct_IO - that is IO requests which bypass the page cache
NeilBrowna9e102b2006-03-25 03:08:29 -0800728 and transfer data directly between the storage and the
NeilBrown341546f2006-03-25 03:07:56 -0800729 application's address space.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700730
731 get_xip_page: called by the VM to translate a block number to a page.
732 The page is valid until the corresponding filesystem is unmounted.
733 Filesystems that want to use execute-in-place (XIP) need to implement
734 it. An example implementation can be found in fs/ext2/xip.c.
735
NeilBrown341546f2006-03-25 03:07:56 -0800736 migrate_page: This is used to compact the physical memory usage.
737 If the VM wants to relocate a page (maybe off a memory card
738 that is signalling imminent failure) it will pass a new page
739 and an old page to this function. migrate_page should
740 transfer any private data across and update any references
741 that it has to the page.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700742
Borislav Petkov422b14c2007-07-15 23:41:43 -0700743 launder_page: Called before freeing a page - it writes back the dirty page. To
744 prevent redirtying the page, it is kept locked during the whole
745 operation.
746
Andi Kleen25718732009-09-16 11:50:13 +0200747 error_remove_page: normally set to generic_error_remove_page if truncation
748 is ok for this address space. Used for memory failure handling.
749 Setting this implies you deal with pages going away under you,
750 unless you have them locked or reference counts increased.
751
752
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800753The File Object
754===============
755
756A file object represents a file opened by a process.
757
758
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700759struct file_operations
Pekka Enbergcc7d1f82005-11-07 01:01:08 -0800760----------------------
Linus Torvalds1da177e2005-04-16 15:20:36 -0700761
762This describes how the VFS can manipulate an open file. As of kernel
Borislav Petkov422b14c2007-07-15 23:41:43 -07007632.6.22, the following members are defined:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700764
765struct file_operations {
Borislav Petkov422b14c2007-07-15 23:41:43 -0700766 struct module *owner;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700767 loff_t (*llseek) (struct file *, loff_t, int);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700768 ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700769 ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
Badari Pulavarty027445c2006-09-30 23:28:46 -0700770 ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
771 ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700772 int (*readdir) (struct file *, void *, filldir_t);
773 unsigned int (*poll) (struct file *, struct poll_table_struct *);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700774 long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
775 long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700776 int (*mmap) (struct file *, struct vm_area_struct *);
777 int (*open) (struct inode *, struct file *);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700778 int (*flush) (struct file *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700779 int (*release) (struct inode *, struct file *);
Josef Bacik02c24a82011-07-16 20:44:56 -0400780 int (*fsync) (struct file *, loff_t, loff_t, int datasync);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700781 int (*aio_fsync) (struct kiocb *, int datasync);
782 int (*fasync) (int, struct file *, int);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700783 int (*lock) (struct file *, int, struct file_lock *);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700784 ssize_t (*readv) (struct file *, const struct iovec *, unsigned long, loff_t *);
785 ssize_t (*writev) (struct file *, const struct iovec *, unsigned long, loff_t *);
786 ssize_t (*sendfile) (struct file *, loff_t *, size_t, read_actor_t, void *);
787 ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
788 unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
789 int (*check_flags)(int);
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700790 int (*flock) (struct file *, int, struct file_lock *);
Borislav Petkov422b14c2007-07-15 23:41:43 -0700791 ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, size_t, unsigned int);
792 ssize_t (*splice_read)(struct file *, struct pipe_inode_info *, size_t, unsigned int);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700793};
794
795Again, all methods are called without any locks being held, unless
796otherwise noted.
797
798 llseek: called when the VFS needs to move the file position index
799
800 read: called by read(2) and related system calls
801
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700802 aio_read: called by io_submit(2) and other asynchronous I/O operations
803
Linus Torvalds1da177e2005-04-16 15:20:36 -0700804 write: called by write(2) and related system calls
805
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700806 aio_write: called by io_submit(2) and other asynchronous I/O operations
807
Linus Torvalds1da177e2005-04-16 15:20:36 -0700808 readdir: called when the VFS needs to read the directory contents
809
810 poll: called by the VFS when a process wants to check if there is
811 activity on this file and (optionally) go to sleep until there
812 is activity. Called by the select(2) and poll(2) system calls
813
Arnd Bergmannb19dd422010-07-04 00:15:10 +0200814 unlocked_ioctl: called by the ioctl(2) system call.
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700815
816 compat_ioctl: called by the ioctl(2) system call when 32 bit system calls
817 are used on 64 bit kernels.
818
Linus Torvalds1da177e2005-04-16 15:20:36 -0700819 mmap: called by the mmap(2) system call
820
821 open: called by the VFS when an inode should be opened. When the VFS
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700822 opens a file, it creates a new "struct file". It then calls the
823 open method for the newly allocated file structure. You might
824 think that the open method really belongs in
825 "struct inode_operations", and you may be right. I think it's
826 done the way it is because it makes filesystems simpler to
827 implement. The open() method is a good place to initialize the
828 "private_data" member in the file structure if you want to point
829 to a device structure
830
831 flush: called by the close(2) system call to flush a file
Linus Torvalds1da177e2005-04-16 15:20:36 -0700832
833 release: called when the last reference to an open file is closed
834
835 fsync: called by the fsync(2) system call
836
837 fasync: called by the fcntl(2) system call when asynchronous
838 (non-blocking) mode is enabled for a file
839
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700840 lock: called by the fcntl(2) system call for F_GETLK, F_SETLK, and F_SETLKW
841 commands
842
843 readv: called by the readv(2) system call
844
845 writev: called by the writev(2) system call
846
847 sendfile: called by the sendfile(2) system call
848
849 get_unmapped_area: called by the mmap(2) system call
850
851 check_flags: called by the fcntl(2) system call for F_SETFL command
852
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700853 flock: called by the flock(2) system call
854
Pekka J Enbergd1195c52006-04-11 14:21:59 +0200855 splice_write: called by the VFS to splice data from a pipe to a file. This
856 method is used by the splice(2) system call
857
858 splice_read: called by the VFS to splice data from file to a pipe. This
859 method is used by the splice(2) system call
860
Linus Torvalds1da177e2005-04-16 15:20:36 -0700861Note that the file operations are implemented by the specific
862filesystem in which the inode resides. When opening a device node
863(character or block special) most filesystems will call special
864support routines in the VFS which will locate the required device
865driver information. These support routines replace the filesystem file
866operations with those for the device driver, and then proceed to call
867the new open() method for the file. This is how opening a device file
868in the filesystem eventually ends up calling the device driver open()
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700869method.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700870
871
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700872Directory Entry Cache (dcache)
873==============================
874
Linus Torvalds1da177e2005-04-16 15:20:36 -0700875
876struct dentry_operations
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700877------------------------
Linus Torvalds1da177e2005-04-16 15:20:36 -0700878
879This describes how a filesystem can overload the standard dentry
880operations. Dentries and the dcache are the domain of the VFS and the
881individual filesystem implementations. Device drivers have no business
882here. These methods may be set to NULL, as they are either optional or
Eric Dumazetc23fbb62007-05-08 00:26:18 -0700883the VFS uses a default. As of kernel 2.6.22, the following members are
Linus Torvalds1da177e2005-04-16 15:20:36 -0700884defined:
885
886struct dentry_operations {
Pekka J Enberg5ea626a2005-09-09 13:10:19 -0700887 int (*d_revalidate)(struct dentry *, struct nameidata *);
Nick Pigginb1e6a012011-01-07 17:49:28 +1100888 int (*d_hash)(const struct dentry *, const struct inode *,
889 struct qstr *);
Nick Piggin621e1552011-01-07 17:49:27 +1100890 int (*d_compare)(const struct dentry *, const struct inode *,
891 const struct dentry *, const struct inode *,
892 unsigned int, const char *, const struct qstr *);
Nick Pigginfe15ce42011-01-07 17:49:23 +1100893 int (*d_delete)(const struct dentry *);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700894 void (*d_release)(struct dentry *);
895 void (*d_iput)(struct dentry *, struct inode *);
Eric Dumazetc23fbb62007-05-08 00:26:18 -0700896 char *(*d_dname)(struct dentry *, char *, int);
David Howells9875cf82011-01-14 18:45:21 +0000897 struct vfsmount *(*d_automount)(struct path *);
Al Viro1aed3e42011-03-18 09:09:02 -0400898 int (*d_manage)(struct dentry *, bool);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700899};
900
901 d_revalidate: called when the VFS needs to revalidate a dentry. This
902 is called whenever a name look-up finds a dentry in the
903 dcache. Most filesystems leave this as NULL, because all their
904 dentries in the dcache are valid
905
Nick Piggin34286d62011-01-07 17:49:57 +1100906 d_revalidate may be called in rcu-walk mode (nd->flags & LOOKUP_RCU).
907 If in rcu-walk mode, the filesystem must revalidate the dentry without
908 blocking or storing to the dentry, d_parent and d_inode should not be
909 used without care (because they can go NULL), instead nd->inode should
910 be used.
911
912 If a situation is encountered that rcu-walk cannot handle, return
913 -ECHILD and it will be called again in ref-walk mode.
914
Nick Piggin621e1552011-01-07 17:49:27 +1100915 d_hash: called when the VFS adds a dentry to the hash table. The first
916 dentry passed to d_hash is the parent directory that the name is
Nick Pigginb1e6a012011-01-07 17:49:28 +1100917 to be hashed into. The inode is the dentry's inode.
918
919 Same locking and synchronisation rules as d_compare regarding
920 what is safe to dereference etc.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700921
Nick Piggin621e1552011-01-07 17:49:27 +1100922 d_compare: called to compare a dentry name with a given name. The first
923 dentry is the parent of the dentry to be compared, the second is
924 the parent's inode, then the dentry and inode (may be NULL) of the
925 child dentry. len and name string are properties of the dentry to be
926 compared. qstr is the name to compare it with.
927
928 Must be constant and idempotent, and should not take locks if
929 possible, and should not or store into the dentry or inodes.
930 Should not dereference pointers outside the dentry or inodes without
931 lots of care (eg. d_parent, d_inode, d_name should not be used).
932
933 However, our vfsmount is pinned, and RCU held, so the dentries and
934 inodes won't disappear, neither will our sb or filesystem module.
935 ->i_sb and ->d_sb may be used.
936
937 It is a tricky calling convention because it needs to be called under
938 "rcu-walk", ie. without any locks or references on things.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700939
Nick Pigginfe15ce42011-01-07 17:49:23 +1100940 d_delete: called when the last reference to a dentry is dropped and the
941 dcache is deciding whether or not to cache it. Return 1 to delete
942 immediately, or 0 to cache the dentry. Default is NULL which means to
943 always cache a reachable dentry. d_delete must be constant and
944 idempotent.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700945
946 d_release: called when a dentry is really deallocated
947
948 d_iput: called when a dentry loses its inode (just prior to its
949 being deallocated). The default when this is NULL is that the
950 VFS calls iput(). If you define this method, you must call
951 iput() yourself
952
Eric Dumazetc23fbb62007-05-08 00:26:18 -0700953 d_dname: called when the pathname of a dentry should be generated.
Matt LaPlanted9195882008-07-25 19:45:33 -0700954 Useful for some pseudo filesystems (sockfs, pipefs, ...) to delay
Eric Dumazetc23fbb62007-05-08 00:26:18 -0700955 pathname generation. (Instead of doing it when dentry is created,
Matt LaPlanted9195882008-07-25 19:45:33 -0700956 it's done only when the path is needed.). Real filesystems probably
Eric Dumazetc23fbb62007-05-08 00:26:18 -0700957 dont want to use it, because their dentries are present in global
958 dcache hash, so their hash should be an invariant. As no lock is
959 held, d_dname() should not try to modify the dentry itself, unless
960 appropriate SMP safety is used. CAUTION : d_path() logic is quite
961 tricky. The correct way to return for example "Hello" is to put it
962 at the end of the buffer, and returns a pointer to the first char.
963 dynamic_dname() helper function is provided to take care of this.
964
David Howells9875cf82011-01-14 18:45:21 +0000965 d_automount: called when an automount dentry is to be traversed (optional).
David Howellsea5b7782011-01-14 19:10:03 +0000966 This should create a new VFS mount record and return the record to the
967 caller. The caller is supplied with a path parameter giving the
968 automount directory to describe the automount target and the parent
969 VFS mount record to provide inheritable mount parameters. NULL should
970 be returned if someone else managed to make the automount first. If
971 the vfsmount creation failed, then an error code should be returned.
972 If -EISDIR is returned, then the directory will be treated as an
973 ordinary directory and returned to pathwalk to continue walking.
974
975 If a vfsmount is returned, the caller will attempt to mount it on the
976 mountpoint and will remove the vfsmount from its expiration list in
977 the case of failure. The vfsmount should be returned with 2 refs on
978 it to prevent automatic expiration - the caller will clean up the
979 additional ref.
David Howells9875cf82011-01-14 18:45:21 +0000980
981 This function is only used if DCACHE_NEED_AUTOMOUNT is set on the
982 dentry. This is set by __d_instantiate() if S_AUTOMOUNT is set on the
983 inode being added.
984
David Howellscc53ce52011-01-14 18:45:26 +0000985 d_manage: called to allow the filesystem to manage the transition from a
986 dentry (optional). This allows autofs, for example, to hold up clients
987 waiting to explore behind a 'mountpoint' whilst letting the daemon go
988 past and construct the subtree there. 0 should be returned to let the
989 calling process continue. -EISDIR can be returned to tell pathwalk to
990 use this directory as an ordinary directory and to ignore anything
991 mounted on it and not to check the automount flag. Any other error
992 code will abort pathwalk completely.
993
David Howellsab909112011-01-14 18:46:51 +0000994 If the 'rcu_walk' parameter is true, then the caller is doing a
995 pathwalk in RCU-walk mode. Sleeping is not permitted in this mode,
996 and the caller can be asked to leave it and call again by returing
997 -ECHILD.
998
David Howellscc53ce52011-01-14 18:45:26 +0000999 This function is only used if DCACHE_MANAGE_TRANSIT is set on the
1000 dentry being transited from.
1001
Eric Dumazetc23fbb62007-05-08 00:26:18 -07001002Example :
1003
1004static char *pipefs_dname(struct dentry *dent, char *buffer, int buflen)
1005{
1006 return dynamic_dname(dentry, buffer, buflen, "pipe:[%lu]",
1007 dentry->d_inode->i_ino);
1008}
1009
Linus Torvalds1da177e2005-04-16 15:20:36 -07001010Each dentry has a pointer to its parent dentry, as well as a hash list
1011of child dentries. Child dentries are basically like files in a
1012directory.
1013
Pekka J Enberg5ea626a2005-09-09 13:10:19 -07001014
Pekka Enbergcc7d1f82005-11-07 01:01:08 -08001015Directory Entry Cache API
Linus Torvalds1da177e2005-04-16 15:20:36 -07001016--------------------------
1017
1018There are a number of functions defined which permit a filesystem to
1019manipulate dentries:
1020
1021 dget: open a new handle for an existing dentry (this just increments
1022 the usage count)
1023
1024 dput: close a handle for a dentry (decrements the usage count). If
Nick Pigginfe15ce42011-01-07 17:49:23 +11001025 the usage count drops to 0, and the dentry is still in its
1026 parent's hash, the "d_delete" method is called to check whether
1027 it should be cached. If it should not be cached, or if the dentry
1028 is not hashed, it is deleted. Otherwise cached dentries are put
1029 into an LRU list to be reclaimed on memory shortage.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001030
1031 d_drop: this unhashes a dentry from its parents hash list. A
Pekka J Enberg5ea626a2005-09-09 13:10:19 -07001032 subsequent call to dput() will deallocate the dentry if its
Linus Torvalds1da177e2005-04-16 15:20:36 -07001033 usage count drops to 0
1034
1035 d_delete: delete a dentry. If there are no other open references to
1036 the dentry then the dentry is turned into a negative dentry
1037 (the d_iput() method is called). If there are other
1038 references, then d_drop() is called instead
1039
1040 d_add: add a dentry to its parents hash list and then calls
1041 d_instantiate()
1042
1043 d_instantiate: add a dentry to the alias hash list for the inode and
1044 updates the "d_inode" member. The "i_count" member in the
1045 inode structure should be set/incremented. If the inode
1046 pointer is NULL, the dentry is called a "negative
1047 dentry". This function is commonly called when an inode is
1048 created for an existing negative dentry
1049
1050 d_lookup: look up a dentry given its parent and path name component
1051 It looks up the child of that given name from the dcache
1052 hash table. If it is found, the reference count is incremented
Zhaoleibe42c4c2008-12-01 14:34:58 -08001053 and the dentry is returned. The caller must use dput()
Linus Torvalds1da177e2005-04-16 15:20:36 -07001054 to free the dentry when it finishes using it.
1055
Miklos Szeredif84e3f52008-02-08 04:21:34 -08001056Mount Options
1057=============
1058
1059Parsing options
1060---------------
1061
1062On mount and remount the filesystem is passed a string containing a
1063comma separated list of mount options. The options can have either of
1064these forms:
1065
1066 option
1067 option=value
1068
1069The <linux/parser.h> header defines an API that helps parse these
1070options. There are plenty of examples on how to use it in existing
1071filesystems.
1072
1073Showing options
1074---------------
1075
1076If a filesystem accepts mount options, it must define show_options()
1077to show all the currently active options. The rules are:
1078
1079 - options MUST be shown which are not default or their values differ
1080 from the default
1081
1082 - options MAY be shown which are enabled by default or have their
1083 default value
1084
1085Options used only internally between a mount helper and the kernel
1086(such as file descriptors), or which only have an effect during the
1087mounting (such as ones controlling the creation of a journal) are exempt
1088from the above rules.
1089
1090The underlying reason for the above rules is to make sure, that a
1091mount can be accurately replicated (e.g. umounting and mounting again)
1092based on the information found in /proc/mounts.
1093
1094A simple method of saving options at mount/remount time and showing
1095them is provided with the save_mount_options() and
1096generic_show_options() helper functions. Please note, that using
1097these may have drawbacks. For more info see header comments for these
1098functions in fs/namespace.c.
Pekka Enbergcc7d1f82005-11-07 01:01:08 -08001099
1100Resources
1101=========
1102
1103(Note some of these resources are not up-to-date with the latest kernel
1104 version.)
1105
1106Creating Linux virtual filesystems. 2002
1107 <http://lwn.net/Articles/13325/>
1108
1109The Linux Virtual File-system Layer by Neil Brown. 1999
1110 <http://www.cse.unsw.edu.au/~neilb/oss/linux-commentary/vfs.html>
1111
1112A tour of the Linux VFS by Michael K. Johnson. 1996
1113 <http://www.tldp.org/LDP/khg/HyperNews/get/fs/vfstour.html>
1114
1115A small trail through the Linux kernel by Andries Brouwer. 2001
1116 <http://www.win.tue.nl/~aeb/linux/vfs/trail.html>